Add testing for L0_iterative_tutorial #7176
Conversation
|
@oandreeva-nv I think we did this for Python backend but the feedback was that it makes it harder to determine which part of the test is flaky since they are all bundled together. Additionally, the tests can be executed in parallel if they are separate. |
qa/L0_iterative_tutorial/test.sh
Outdated
| @@ -0,0 +1,49 @@ | |||
| # Copyright (c) 2024, NVIDIA CORPORATION & AFFILIATES. All rights reserved. | |||
There was a problem hiding this comment.
nit similar to Olga's comment - can we rename this to L0_tutorial_iterative_scheduling (and future L0_tutorial_*) so we can quickly identify/group the tutorial repo based tests?
Similar to L0_backend_{identity,python,...}
| RET=0 | ||
|
|
||
| # Install the dependencies | ||
| pip3 install "transformers[torch]" |
There was a problem hiding this comment.
Should we lock down the version to whatever worked while testing to avoid unexpected breakage?
| # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. | ||
|
|
||
| rm -rf tutorials | ||
| git clone https://github.com/triton-inference-server/tutorials.git -b ${TUTORIAL_REPO_TAG} --depth 1 |
There was a problem hiding this comment.
Should we default TUTORIAL_REPO_TAG for ergonomics of local testing (docker exec devel) outside of a pipeline? Something like this
TUTORIAL_REPO_TAG=${TUTORIAL_REPO_TAG:-"main"}
triton-inference-server/tutorials#90